Victor Steinberg Video Standards Signals, Formats and Interfaces Part 6 Color Spaces, Formats, Levels, Ranges www. videoq. com 1
Analog & Digital, RGB & YUV, Component & Composite Some formats shown on this diagram have a local variants, e. g. NTSC-USA uses the 7. 5 IRE setup, but NTSC-Japan does not; also there are variants restricted to particular scanning standard, e. g. SECAM exists only in 625 i 25 format, whilst for PAL there are several variants. As to analog and digital levels, there are not so many variants. For analog signals there are two schemes, the main difference is the signal range between Reference Black and Reference White: 700 m. V in 50 Hz countries 100 IRE (714. 285714 m. V) in 60 Hz countries Many years ago, the black and white video signal nominal peak-to-peak level was set at 1 V (1000 m. V) - a simple, elegant choice. However, the Sync to Luminance ratio was different: 4 to 10 in America and 3 to 7 in other countries. This difference resulted in a discrepancy between nominal (reference) white levels, i. e. the voltage change needed to go from black to white. In 50 Hz countries it is 700 m. V, but in America it is 10000/14 = 714. 285714 m. V. 100 IRE 700 m. V The latter figure does not exactly roll off the tongue, and so the Institute of Radio Engineers (IRE) devised a special unit of measurement: the "IRE". 1 IRE is defined as 7. 14 m. V; the levels range from black to white is 100 IRE, and the luminance plus sync peak-to-peak range is 140 IRE. 40 IRE 300 m. V 14 m. V © 2017 Video. Q, Inc. www. videoq. com 2
Color Bars – Checking RGB & YUV Levels Color Bars are the most common pattern for testing video ADCs, DACs, RGB YUV conversion, encoders, decoders, codecs, displays, etc. For a start, let’s look at the analog component interface color bars levels. The simplest and oldest color bars variant is a sequence of eight vertical bars of 100% (maximum intensity) colors: White, Yellow, Cyan, Green, Magenta, Blue, Black. Analog HD & SD Y Pb (U) YPb. Pr interface Pr (V) 700 m. V R 100% G +/- 50% 350 m. V Pb (U) Y BT. 709 (HD) Pr (V) BT. 601 (SD) B Analog HD & SD RGB interface No analog RGB or YPb. Pr interface for UHD, only digital Video. Q VQV: In Vector. Scope all HD and SD UV vectors hit the 100% Bars target boxes. YUV waveforms and the target boxes positions are different, but peak-to-peak HD & SD ranges are equal. © 2017 Video. Q, Inc. www. videoq. com 3
Digital RGB & YUV Ranges In broadcast environment Y, R, G and B 8 bit values are conventionally shifted and scaled to the range [16, 235] referred to as “Narrow Range” = “NR” (aka “Limited Range”, “Low RGB”, "TV“, “Broadcast”, etc). Thus, the digital 8 bit Reference White Level is 235 (940 on 10 bit scale) and the 8 bit Reference Black Level is 16 (64 on 10 bit scale). The headroom above 235 and the footroom below 16 accommodate signal overshoots ("ringing") due to filtering and specular highlights. The [0, 255] range referred to as “Full Range” = “FR” (aka “High RGB”, "PC“, “CG”, etc) is also widely used for digital RGB interfaces, e. g. HDMI. For YUV signals the Narrow Range ( [16, 235 Y], [16, 240 UV]) is mandatory; some consumer product use non-standard ”yuvj” [0, 255] color space. In broadcast environment levels from 1 to 254 are available for video, levels 0 and 255 are used exclusively for SDI interface synchronization. In file based environment all levels from 0 to 255 are available for video. Narrow Range Levels Mapping Headroom = 19 Range = 219 Footroom = 15 Headroom = 14 Y, R, G, B Range = 224 U V Offset = 128 Footroom = 15 Offset = 16 © 2017 Video. Q, Inc. www. videoq. com 4
RGB YUV Conversion Matrices Fundamental RGB to Y conversion coefficients are defined by the international standards: BT. 601 (SD) BT. 709 (HD) BT. 2020 -NCL (UHD-SDR), BT. 2100 -NCL (UHD-HDR & HD-HDR) 1. All matrices are balanced, i. e. the sums of their coefficients = 1, so the R = G = B = 1 (100% white) input produces Y = 1 (max Y value) output. 2. The Y signal produced by these matrices has not much in common with the CIE 1931 Luminance Y. It is a “codec Y” re-using the same letter Y. The choice of these RGB to Y matrices coefficients was not based on the colorimetry factors, but on the content delivery quality factors. 3. The RGB YUV matrices are just derivatives of the 3 fundamental column matrices as above; the unity range principle remains, i. e. on black Y = R = G = B = 0 and U = V = 0, on white Y = R = G = B = 1 and U = V = 0, Y range is [0, 1] and UV range is [-0. 5, +0. 5]: © 2017 Video. Q, Inc. www. videoq. com 5
Full Range Narrow Range – Digital Levels Mapping RGB YUV Conversion Standards define the matrix coefficients in relative units, e. g. [0, 1] or [0%, 100%]. Digital signals are usually defined not in percents, but in 8, 10 or 12 bit levels. Full Narrow Range Conversion Control Also there two types of digital signal ranges in use: Full & Narrow. Y & UV Ranges Gain Controls RGB to YUV Color Space Y Offset (switchable) Conversion Matrix U & V Offsets (fixed) For example, whilst mapping the [0%, 100%] range to 8, 10 or 12 bit Narrow Range, the YUV offsets and gains look like this: where BAF is Bit Accuracy Factor, function of Bit Depth = 8, 10, 12, etc. bits B G Full to Narrow digital video signal range mapping is the recommended default for RGB to YUV conversion in file based environment. R For the inverse YUV to RGB conversion the recommended default is obviously Narrow to Full range mapping. x x x + + + Y U V But, in practice, there also other RGB/YUV conversion schemes in use, e. g. : For any bit depth the RGB to YUV gain factors are: - Full YUV to Full RGB Full to Narrow conversion (shrinking the range, default): Y mapping gain = 219/255, UV mapping gain = 224/255 Narrow to Full conversion (expanding the range, seldom used): Y mapping gain = 255/219, UV mapping gain = 255/224 - Narrow YUV to Narrow RGB - Narrow RGB to Narrow YUV © 2017 Video. Q, Inc. www. videoq. com 6
RGB & YUV Conversion Variants Source: Destination YUV NR default YUV FR seldom used, e. g. “yuvj” RGB NR default RGB FR used quite often COLOR SPACE CONVERSION NOT NEEDED RANGE SHRINKING RANGE PRESERVATION RANGE SHRINKING RANGE EXPANSION COLOR SPACE CONVERSION NOT NEEDED RANGE EXPANSION RANGE PRESERVATION RANGE SHRINKING COLOR SPACE CONVERSION NOT NEEDED RANGE SHRINKING RANGE EXPANSION RANGE PRESERVATION RANGE EXPANSION COLOR SPACE CONVERSION NOT NEEDED © 2017 Video. Q, Inc. www. videoq. com 7
YUV RGB – Selecting Conversion Mode In the ideal world, the Source Color Space (Color Matrix and YUV Range used upstream) is specified in the embedded metadata. If so, the AUTO controller takes care about all YUV to RGB conversion parameters. But in real world the metadata may be wrong or missing. In such case the “safe default” AUTO approach may help. E. g. if the frame height is smaller than 600 and the aspect ratio is about 4: 3, then the assumed matrix should be BT. 601. If HD originated video was down-scaled (without color matrix conversion) to the anamorphic 720 x 576 frame size for Internet distribution, i. e. to the (strictly speaking illegal) combination of small frame size and HDTV BT. 709 color space, then only a QA/QC operator manual intervention may help. Video. Q VQV Color Matrix selection menu: Only in Band #1 of this test all colors are correct. Video. Q VQV Vector. Scope shows how far away UV vectors may go in case of incorrect selection of Color Matrix and Digital Range. © 2017 Video. Q, Inc. www. videoq. com 8
YUV Color Space Problems Auto Detection Video. Q VQMA Analyzer in 2 seconds checks, among other parameters, all color values within the VQMA Test Pattern and produces: - Machine-readable Report for robots - Visual Report for QA/QC operator Somewhere in the long processing chain the original UHD color space BT. 2020 was converted by mistake to BT. 601. But, according to the UHD frame size, it should be BT. 2020. VQMA spotted the error and produced a failure Report for robots as well as the warning message for QA/QC operator. © 2017 Video. Q, Inc. www. videoq. com 9
About This Presentation Produced by Josef Marc Written by Victor Steinberg, Ph. D Narrated by Josef Marc Conceived by Roderick Snell Technical consulting by Maxim Levkov Based on the book "Video Standards: Signals, Formats and Interfaces" by Victor Steinberg Published by Snell & Wilcox For further reading we recommend wikipedia. org © 2017 Video. Q, Inc. www. videoq. com 10
About Video. Q Company History • Founded in 2005 • Formed by an Engineering Awards winning team sharing between them decades of global video technology. • Video. Q is a renowned player in calibration and benchmarking of video processors, transcoders and displays, providing tools and technologies instantly revealing artifacts, problems and deficiencies, thus raising the bar in productivity and video quality experience. • Video. Q products and services cover all aspects of video processing and quality assurance - from visual picture quality estimation and quality control to fully automated processing, utilizing advanced Video. Q algorithms and robotic video quality analyzers, including latest UHD and HDR developments. Operations • Headquarters in Sunnyvale, CA, USA • Software developers in Silicon Valley and worldwide • Distributors and partners in several countries • Sales & support offices in USA, UK © 2017 Video. Q, Inc. www. videoq. com 11
Скачать презентацию Victor Steinberg Video Standards Signals Formats and Interfaces
4491392d7083aa837a25f1eb7e83ff54.ppt